Table 2.

Diagnostic performances of different models across tenfold cross-validation

Models	Accuracy	Sensitivity	Specificity	PPV	NPV	AUC
Peritumoral	0.724 ± 0.113	0.786 ± 0.134	0.665 ± 0.158	0.723 ± 0.150	0.740 ± 0.150	0.789 ± 0.099
Intratumoral	0.650 ± 0.125	0.738 ± 0.158	0.558 ± 0.126	0.652 ± 0.138	0.658 ± 0.194	0.733 ± 0.092
Original	0.669 ± 0.097	0.664 ± 0.166	0.708 ± 0.154	0.716 ± 0.166	0.656 ± 0.143	0.742 ± 0.079
Intratumoral + peritumoral	0.703 ± 0.125	0.703 ± 0.125	0.703 ± 0.125	0.703 ± 0.125	0.703 ± 0.125	0.803 ± 0.099
Clinical	0.719 ± 0.100	0.739 ± 0.128	0.696 ± 0.170	0.744 ± 0.114	0.715 ± 0.136	0.794 ± 0.081
Ensemble	0.756 ± 0.070	0.786 ± 0.135	0.717 ± 0.132	0.773 ± 0.091	0.763 ± 0.115	0.844 ± 0.063

The results of tenfold cross-validation are expressed as mean ± standard deviation

PPV positive predictive value, NPV negative predictive value, AUC area under the curve, CI confidence interval